RedFromMI wrote: ↑Fri May 15, 2020 2:34 pm
wgdsr wrote: ↑Fri May 15, 2020 2:08 pm
RedFromMI wrote: ↑Fri May 15, 2020 12:54 pm
6ftstick wrote: ↑Fri May 15, 2020 12:40 pm
MDlaxfan76 wrote: ↑Fri May 15, 2020 12:31 pm
6ftstick wrote: ↑Fri May 15, 2020 12:30 pm
MDlaxfan76 wrote: ↑Fri May 15, 2020 12:20 pm
calourie wrote: ↑Thu May 14, 2020 11:42 am
The way you use the numbers has nothing to do with deaths per million. I'm not sure know how to explain it more simply, but I'll try. if you reduce 328 million by a factor of 328 to get it down to one million, you have to reduce the 87,000 by a factor of 328 to find the number per million. Reducing by a factor of 328 requires dividing by 328. 87000 divided by 328 equals 257.
painful isn't it?
It appears a waste of time anyway. The death totals are inflated and the number infected is exponentially greater.
Epidemiologist 6ft, who can't do simple math, speaks...
There you are living in a vacuum again.
The death totals are not inflated - they are lower than reality. If you disagree with that statement you disagree with all the experts.
The statement that the number infected is exponentially greater is pretty inexact, but the general usage of that phrase would mean much larger. False.
The place in the US with the highest infection rate is NYC - with somewhere just over 20% of the population infected. (And given the death tolls including an estimate of the excess deaths you get a death rate of around .9%. Probably is a bit lower than that but not near the .1% associated with a typical flu season.)
Most parts of the US the infection rate is MUCH lower. For Florida, the estimated infection rate is about 1.3%. For South Carolina, where I live, it is about 1.1%. For Michigan (another state with a lot of early cases) it is about 5.9%.
in almost all cases of epidemics that i've seen, estimates do get marked up after the fact. how they do that, imprecise or precise, i don't know and it probably varies.
has nyc or nys done further anti-body tests than the statewide one they released several weeks or a month or so ago? where are florida and south carolina and michigan numbers coming from, if you have them?
The other state numbers are estimates from covid19-projections.com - you would have to check out their references as to how they get their numbers, but their estimates of everything else seems pretty good. I prefer their method over the UW group stuff.
As to calculating real death tolls - you compare actual deaths in a particular week with multiyear averages and calculate the excess deaths compared to "normal." Current estimates for the US seem to be that the actual numbers compared to the official tolls are roughly 1.4-1.5 times bigger. In some other countries, the excess number is a higher percentage increase, but in the US we are starting to get a handle on the numbers, and some states like NY are already trying to get closer to actual (even without a positive test result).
As far as excess death calculations, NY Times and talkingpointsmemo.com have had some good articles.
there will be plenty of deaths missed that maybe get confirmed via blood tests at later dates, particularly as testing is more readily available. there will also be deaths that were exacerbated by the pandemic and measures taken. they will be covid related for sure, but not as a result of covid causing the death by disease. there's a lot of tabulating to do.
https://www.miamiherald.com/news/corona ... 60406.html
not sure what that covid website is using, but this is an actual serological antibody testing survey done, with the results posted from april 24th, over 20 days ago.
1400 participants, they had residents of miami-dade at 6% ish, or approx 165,000 when, on april 24, they had 10,600 cases in the county. now there are 14,570 cases, a 37% increase. back of the napkin, that would mean 15x the number of actual cases vs confirmed cases, and by now dade county would account for an over 1% infection rate all by themselves.
there are now 43,210 confirmed cases in florida, which given the constraints of extrapolating, gives the state an infection estimate of over 3%.
that serological testing was also done over 2 weeks, a period where the county 2 weeks prior to the 4/24 10,600 number actually had about 6,000 confirmed cases. meaning the numbers could've been higher just from the length of time they were conducting tests. there's a link in the article that says UM will continue to do weekly tests, 750 per week. i'm guessing we'll be seeing a lot more of these studies pop up in the interim of time before we start getting the tens or 100s of millions of serological tests ready in the coming months.